refactor(rivetkit-core): remove prevent_sleep field and deprecate set_prevent_sleep to no-op

[NathanFlurry]

Description

Please include a summary of the changes and the related issue. Please also include relevant motivation and context.

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How Has This Been Tested?

Please describe the tests that you ran to verify your changes.

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[NathanFlurry]

• chore(kitchen-sink): configurable endpoint #4766 : 3 dependent PRs (#4754 , #4777 , #4778 )
• refactor(rivetkit-core): rename StopReason to ShutdownKind and kill enum fallthroughs #4765
• chore(envoy-client): log command and event payloads #4751
• fix(rivetkit): minor sleep-cleanup follow-ups #4761
• docs(rivetkit): clean up stale setPreventSleep references in docs and JSDoc #4760
• docs(rivetkit): document keepAwake/waitUntil counter-arm race and surface bridge errors #4759
• test(engine): force overdue branch in alarm-during-sleep test with negative offset #4758
• fix(rivetkit-core): return stopping not starting when sleep/destroy called mid-shutdown #4757
• test(rivetkit): remove sleepWithPreventSleep fixture and tests #4756
• test(engine): add regression test for alarm-during-sleep-transition race #4755
• test(rivetkit-core): cover sleep/destroy guards and prevent_sleep no-op #4750
• docs: add keepAwake vs waitUntil table and internal sleep-sequence invariants #4749
• feat(rivetkit): expose c.keepAwake(promise) through native path #4748
• refactor(rivetkit): deprecate setPreventSleep to no-op in TypeScript #4747
• fix(rivetkit-core): enumerate non-drained counters in grace deadline warn log #4746
• fix(rivetkit-core): abort graceful cleanup task when grace deadline elapses #4745
• fix(rivetkit-core): error on sleep/destroy before startup or already-requested #4743
• fix(rivetkit-core): cap SerializeState shutdown timeout at 15s #4742
• refactor(rivetkit-core): remove prevent_sleep field and deprecate set_prevent_sleep to no-op #4741 👈 (View in Graphite)
• fix(api-public): remove orphan import_export actors module #4764
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[claude]

PR Review: Remove prevent_sleep and Deprecate to No-op

Overview

This PR removes the prevent_sleep field and its associated AtomicBool/Notify machinery from rivetkit-core, replacing set_prevent_sleep and prevent_sleep with deprecated no-ops. The Rust-side cleanup is mechanically correct, but there is a behavioral concern on the native/agentOS actor path worth addressing before merge.

---

Critical: agentOS syncPreventSleep becomes a silent no-op on the native path

The agentOS actor factory (rivetkit-typescript/packages/rivetkit/src/agent-os/actor/index.ts) uses syncPreventSleep(c) to keep actors alive while sessions, processes, hooks, and shells are active. This calls c.setPreventSleep(...), which on the native actor path flows through NativeActorContextAdapter.setPreventSleep() in native.ts, which now delegates to the Rust no-op.

The consequence: agentOS actors with active sessions, processes, or shells can now be put to sleep mid-work. The agentOS sleepGracePeriod is 900 seconds, so this may not manifest immediately in practice, but the invariant that "actor stays awake while activeSessionIds.size > 0 || activeProcesses.size > 0" is silently broken.

The agentOS path needs to be migrated to keep_awake / wait_until before this PR removes prevent_sleep from being effective. Session and process lifecycle hooks should hold a keepAwake guard for the duration of active work instead of calling syncPreventSleep.

Required fix before merge: Either migrate syncPreventSleep in agentOS to a keepAwake-guard-based approach, or gate this change so it only becomes effective for agentOS once that migration is done. Otherwise, agentOS actors may sleep with active sessions.

---

Minor Issues

Stale invariant docs. Any comments or documentation listing prevent_sleep as a can_sleep input should be updated to remove it, to avoid confusing future contributors.

Unused parameter naming. In rivetkit/src/context.rs, the delegating set_prevent_sleep(enabled: bool) passes enabled to the inner no-op. Since the inner function ignores it, the outer parameter could be _enabled for consistency.

Consider #[doc(hidden)] on deprecated methods. Since set_prevent_sleep and prevent_sleep always no-op/return false and are kept only for NAPI bridge compatibility, annotating them with #[doc(hidden)] alongside #[deprecated] would keep them out of rustdoc while still compiling.

---

What Looks Good

• The AtomicBool removal and prevent_sleep_notify: Arc<Notify> removal from the registry are clean with no dangling references.
• The CanSleep::PreventSleep variant is fully removed from both the enum definition and all match arms in sleep.rs.
• The #[allow(deprecated)] annotations in rivetkit-napi and rivetkit/context.rs correctly suppress call-site warnings while preserving the bridge surface.
• The #[deprecated] annotation on both crates will surface warnings to downstream Rust consumers, which is the right signal.
• The can_finalize_sleep predicate removal is consistent with the can_arm_sleep_timer removal — no half-removed guard conditions remain.
• The wait_for_shutdown_tasks drain loop correctly drops the !self.prevent_sleep() guard since wait_until tasks now track via shutdown_tasks and keep_awake guards cover the other cases.

---

Summary

The Rust-side removal is mechanically sound. The main concern is the behavioral regression for agentOS actors where syncPreventSleep calls c.setPreventSleep() on the native adapter, which is now a no-op. This should be addressed via a keepAwake-guard migration in agentOS before landing, or the agentOS path should be explicitly handled so actors don't sleep with active sessions.